# AI Ethics **AI ethics** is the body of principles, institutions and fights over how artificial-intelligence systems should be built, deployed and governed — [[Algorithmic Bias]], surveillance, labor displacement, synthetic media, and the question of whose values a model encodes. It is in the vault because the field itself became a front in the culture war: the people who built it were fired, the institutions that housed it were defunded, and the administration that took office in 2025 wrote its abolition into policy under the name of neutrality. ## Two fields with one name There are two AI-ethics traditions and they do not like each other. The first — associated with [[Timnit Gebru]], [[Emily Bender]] and the "Stochastic Parrots" paper (2021) — studies present harms: biased outputs, labor exploitation in data work, environmental cost, and the concentration of power in a few labs; its founding episode is the [[Google AI Ethics Firing]] of Gebru in December 2020. The second — descended from [[Effective Altruism]] and the [[Existential Risk]] literature — studies hypothetical future harms from superintelligent systems, and is funded by the labs whose products the first tradition criticizes. The right attacks the first as "woke" and, increasingly, courts the second as an ally against regulation of the labs it now shares donors with. See [[TESCREAL]] for the second tradition's genealogy. ## The right's position Conservative AI politics inverted between 2020 and 2025. The first phase was grievance: that content-moderation and safety systems suppressed conservative speech, that [[Big Tech]]'s models embedded progressive values — the "woke AI" critique that [[Elon Musk]] used to launch xAI and Grok (see [[Algorithmic Bias]] and [[Censorship]]). The second phase was governance: the [[AI Action Plan]] (July 23, 2025) and its *Preventing Woke AI in the Federal Government* order made "ideological neutrality" a procurement requirement, the December 2025 preemption order set the Justice Department against state AI laws, and the June 2, 2026 frontier-model order guaranteed in writing that no licensing regime would follow. The field that existed to constrain the labs was, in policy terms, replaced by a field that exists to constrain the field. The tension inside the right is real: a populist wing ([[Steve Bannon]], [[Josh Hawley]], the state attorneys general who killed the 2025 moratorium on state AI laws 99–1) wants the labs constrained; the tech wing in the White House wants them unconstrained. Deepfakes are where the two meet — the right is both a producer and a target of synthetic [[Disinformation]], and opposes most regulation of it as a speech-control vehicle. ## Why it matters here AI ethics is the clearest case in the vault of a technical discipline converted into a partisan marker. The substantive questions — who is harmed by a hiring model, who owns the data, what a model may refuse to say — did not go away; they were reassigned. See Bioethics for the earlier precedent, [[Artificial Intelligence]] for the technology, and [[Content Moderation]] for the adjacent fight. ## See Also - [[Artificial Intelligence]] · [[Algorithmic Bias]] · [[Big Tech]] · [[Censorship]] · [[Disinformation]] · [[Content Moderation]] - [[Timnit Gebru]] · [[Emily Bender]] · [[Google AI Ethics Firing]] · [[Effective Altruism]] · [[Existential Risk]] · [[TESCREAL]] - [[AI Action Plan]] · [[Elon Musk]] · Bioethics